SlideShare une entreprise Scribd logo
1  sur  17
Télécharger pour lire hors ligne
DB2 pureScale Performance
Steve Rees
srees@ca.ibm.com Oct 19, 2010
2
Copyright IBM 2010
Agenda
DB2 pureScale technology review
RDMA and low-latency interconnect
Monitoring and tuning bufferpools
in pureScale
Architectural features for top
performance
Performance metrics
3
Copyright IBM 2010
Cluster Interconnect
DB2 pureScale : Technology Review
Single Database View
Clients
Database
Log Log Log Log
Shared Storage Access
CS CS CSCS
CS CS
CS
Member Member Member Member
Primary2nd-ary
DB2 engine runs on several host computers
– Co-operate with each other to provide coherent access to the
database from any member
Data sharing architecture
– Shared access to database
– Members write to their own logs
– Logs accessible from another host (used during recovery)
Cluster Caching Facility (CF) technology from STG
– Efficient global locking and buffer management
– Synchronous duplexing to secondary ensures availability
Low latency, high speed interconnect
– Special optimizations provide significant advantages on RDMA-
capable interconnects like Infiniband
Clients connect anywhere,…
… see single database
– Clients connect into any member
– Automatic load balancing and client reroute may change
underlying physical member to which client is connected
Integrated cluster services
– Failure detection, recovery automation, cluster file system
– In partnership with STG (GPFS,RSCT) and Tivoli (SA MP)
Leverage IBM’s System z Sysplex Experience and Know-How
4
Copyright IBM 2010
DB2 pureScale and low-latency interconnect
Infiniband & uDAPL provide the
low-latency RDMA infrastructure
exploited by pureScale
pureScale currently uses
DDR and QDR IB adapters
according to platform
– Peak throughput of about
2-4 M messages per second
– Provide message latencies
in the 10s of microseconds
or even lower
The Infiniband development roadmap indicates continued increases in bit rates
Infiniband Roadmap from www.infinibandta.org
5
Copyright IBM 2010
Two-level page buffering – data consistency & improved performance
The local bufferpool (LBP) caches both read-only
and updated pages for that member
The shared GBP contains references to every page
in all LBPs across the cluster
– References ensure consistency across
members – who’s interested in which pages, in
case the pages are updated
The GBP also contains copies of all updated pages
from the LBPs
– Sent from the LBP at transaction commit time
– Stored in the GBP & available to members on
demand
– 30 µs page read request over Infiniband from
the GBP can be more than 100x faster than
reading from disk
Statistics are kept for tuning
– Found in LBP vs. found in GBP vs. read
from disk
– Useful in tuning GBP / LBP sizes
CF
M1 M2 M3
10µs
3
5000µs
60µs
30
µs
2
30µs
1
Expensive disk
reads from M1,
M2 not required –
get the modified
page from the CF
#1
#2
#3
#4#5
6
Copyright IBM 2010
pureScale bufferpool monitoring and tuning
Familiar DB2 hit ratio calculations are useful with pureScale
– HR = (logical reads – physical reads) / logical reads
e.g. (pool_data_l_reads – pool_data_p_reads)/pool_data_l_reads
– As usual, physical reads come from disk, logical reads from the bufferpool
(in pureScale, either this means either the LBP or GBP)
e.g., pool_data_l_reads = pool_data_lbp_pages_found +
pool_data_gbp_l_reads
New metrics in pureScale support breaking this down by LBP & GBP amounts
– pool_data_lbp_pages_found = logical data reads satisfied by the LBP
• i.e., we needed a page, and it was present & valid in the LBP
– pool_data_gbp_l_reads = logical data reads attempted at the GBP
• i.e., either not present or not valid in the LBP, so we needed to go
to the GBP
– pool_data_gbp_p_reads = physical data reads due to page not present in
either the LBP or GBP
• Essentially the same as non-pureScale pool_data_p_reads
– pool_data_gbp_invalid_pages = number of GBP data read attempts due
to an LBP page being present but marked invalid
• An indicator of the rate of GBP updates & their impact on the LBP
Of course,
there are index
ones too
7
Copyright IBM 2010
pureScale bufferpool monitoring
Overall (and non-pureScale) hit ratio
– (pool_data_l_reads – pool_data_p_reads)/pool_data_l_reads
– Great values: 95% for index, 90% for data
– Good values: 80-90% for index, 75-85% for data
LBP hit ratio
– (pool_data_lbp_pages_found / pool_data_l_reads) * 100%
– Generally lower than the overall hit ratio, since it excludes GBP hits
– Factors which may affect it, other than LBP size
• Increases with greater portion of read activity in the system
– Decreasing probability that LBP copies of the page have been invalidated
• May decrease with cluster size
– Increasing probability that another member has invalidated the LBP page
GBP hit ratio
– (pool_data_gbp_l_reads – pool_data_gbp_p_reads) /
pool_data_gbp_l_reads
– A hit here is a read of a previously modified page, so hit ratios are typically quite low
• An overall (LBP+GBP) H/R in the high 90's can correspond to a GBP H/R in the
low 80's
– Factors which may affect it, other than GBP size
• Decreases with greater portion of read activity
8
Copyright IBM 2010
pureScale bufferpool tuning
Step 1: typical rule-of-thumb for GBP size = 35-40% of Σ( all members’ LBP sizes )
e.g. 4 members, LBP size of 1M pages each -> GBP size of 1.4 to 1.6M pages
NB - don't forget, GBP page size is always 4kB, no matter what the LBP page size is.
– If your workload very read-heavy (e.g. 90% read), initial GBP allocation could be in the
20-30% range
– For 2-member clusters, you may want to start with 40-50% of total LBP, vs. 35-40%
Step 2: monitor the overall BP hit ratio as usual, with pool_data_l_reads,
pool_data_p_reads, etc.
– Meets your goals? If yes, then done!
Step 3: check LBP H/R with pool_data_lbp_pages_found/pool_data_l_reads
– Great values: 90% for index, 85% for data
– Good values: 70-80% for index, 65-80% for data
– Increasing LBP size can help increase LBP H/R
– NB – for each 16 extra LBP pages, the GBP needs 1 extra page for registrations
Step 4: check GBP H/R with pool_data_gbp_l_reads, pool_data_gbp_p_reads, etc.
– Great values: 90% for index, 80% for data
– Good values: 65-80% for index, 60-75% for data
– pool_data_l_reads > 10 x pool_data_gbp_l_reads means low GBP
dependence – may mean tuning GBP size in this case is less valuable
– pool_data_gbp_invalid_pages > 25% of pool_data_gbp_l_reads means
GBP is really helping out, and could benefit from extra pages
9
Copyright IBM 2010
Page lock negotiation – or Psst! Hey buddy, can you pass me that page?
– pureScale page locks are physical locks, indicating which member currently ‘owns’
the page. Picture the following:
• Member A : acquires a page P and modifies a row on it, and continues with its
transaction. ‘A’ holds an exclusive page lock on page P until ‘A’ commits
• Member B : wants to modify a different row on the same page P. What now?
– ‘B’ doesn’t have to wait until ‘A’ commits & releases the page lock
• The CF will negotiate the page back from ‘A’ in the middle of ‘A’s transaction,
on ‘B’s behalf
• Provides far better concurrency & performance than needing to wait for a page
lock until the holder commits.
Log P
P
pureScale architectural features for optimum performance
P P
Member A Member B
Log
P ?P !
CF
GLM
Px
: A: B
10
Copyright IBM 2010
pureScale architectural features for optimum performance
Table append cache and index page cache
– What happens in the case of rapid inserts into a single table by
multiple members? Or rapid index updates?
Will it cause the insert page to ‘thrash’ back & forth between the
members, each time one has a new row?
– No - each member sets aside an extent for insertion into the table to
eliminate contention & page thrashing. Similarly for indexes with the
page cache
Lock avoidance
– pureScale exploits cursor stability (CS) locking semantics to
avoid taking locks in many common cases
– Reduces pathlength and saves trips to the CF
– Transparent & always on
11
Copyright IBM 2010
Notes on storage configuration for performance
GPFS best practices
– Automatically configured by db2cluster command
• Blocksize >= 1 MB (vs. default 64k) provides
noticeably improved performance
• Direct (unbuffered) IO for both logs & tablespace
containers
• SCSI-3 P/R on AIX enables faster disk takeover on
member failure
– Separate paths for logs & tablespaces are recommended
Dominant storage performance factor for pureScale: fast
log writes
– Always important in OLTP
– Extra important in pureScale due to log flushes driven
by page reclaims
– Separate filesystems, separate devices from each
other & from tablespaces
– Ideally – comfortably under 1ms
– Possibly even SSDs to keep write latencies as
low as possible
12
Copyright IBM 2010
12 Member Scalability Example
Moderately heavy transaction processing
workload modeling warehouse & ordering
process
– Write transactions rate 20%
– Typical read/write ratio of many OLTP
workloads
No cluster awareness in the app
– No affinity
– No partitioning
– No routing of transactions to members
Configuration
– Twelve 8-core p550 members, 64 GB, 5 GHz
– IBM 20Gb/s IB HCAs + 7874-024 IB Switch
– Duplexed PowerHA pureScale across 2 additional
8-core p550s, 64 GB, 5 GHz
– DS8300 storage 576 15K disks, Two 4Gb FC Switches
1Gb Ethernet
Client
Connectivity
20Gb IB
pureScale
Interconnect
7874-024
Switch
Two 4Gb FC
Switches
DS8300
Storage
p550
members
p550
Cluster Caching Facility
Clients (2-way x345)
13
Copyright IBM 2010
12 Member Scalability Example - Results
0
1
2
3
4
5
6
7
8
9
10
11
12
0 5 10 15
1.98x @ 2 members
3.9x @ 4 members
# Members
Throughputvs.1member
7.6x @ 8 members
10.4x @ 12 members
14
Copyright IBM 2010
DB2 pureScale Architecture Scalability
How far will it scale?
Take a web commerce type workload
– Read mostly but not read only – about 90/10
Don’t make the application cluster aware
– No routing of transactions to members
– Demonstrate transparent application scaling
Scale out to the 128 member limit and measure scalability
15
Copyright IBM 2010
The 128-member result
64 Members
95% Scalability
16 Members
Over 95%
Scalability
2, 4 and 8
Members Over
95% Scalability
32 Members
Over 95%
Scalability
88 Members
90% Scalability
112 Members
89% Scalability
128 Members
84% Scalability
16
Copyright IBM 2010
Summary
Performance & scalability are two top goals of
pureScale
– many architectural features were designed
solely to drive the best possible performance
Monitoring and tuning for pureScale extends
existing DB2 interfaces and practices
– e.g., techniques for optimizing GBP/LBP
configuration builds on steps already familiar
to DB2 DBAs.
The pureScale architecture exploits leading-edge
low latency interconnects and RDMA, to achieve
excellent performance & scalability
– Initial 12- & 128-member proofpoints are
strong evidence of a successful first release,
with even better things to come!
17
Copyright IBM 2010
Questions

Contenu connexe

Tendances

DB2 for z/OS Architecture in Nutshell
DB2 for z/OS Architecture in NutshellDB2 for z/OS Architecture in Nutshell
DB2 for z/OS Architecture in NutshellCuneyt Goksu
 
DB2 Accounting Reporting
DB2  Accounting ReportingDB2  Accounting Reporting
DB2 Accounting ReportingJohn Campbell
 
DB2 11 for z/OS Migration Planning and Early Customer Experiences
DB2 11 for z/OS Migration Planning and Early Customer ExperiencesDB2 11 for z/OS Migration Planning and Early Customer Experiences
DB2 11 for z/OS Migration Planning and Early Customer ExperiencesJohn Campbell
 
Db2 recovery IDUG EMEA 2013
Db2 recovery IDUG EMEA 2013Db2 recovery IDUG EMEA 2013
Db2 recovery IDUG EMEA 2013Dale McInnis
 
DB2 for z/OS Real Storage Monitoring, Control and Planning
DB2 for z/OS Real Storage Monitoring, Control and PlanningDB2 for z/OS Real Storage Monitoring, Control and Planning
DB2 for z/OS Real Storage Monitoring, Control and PlanningJohn Campbell
 
DB2 for z/OS Bufferpool Tuning win by Divide and Conquer or Lose by Multiply ...
DB2 for z/OS Bufferpool Tuning win by Divide and Conquer or Lose by Multiply ...DB2 for z/OS Bufferpool Tuning win by Divide and Conquer or Lose by Multiply ...
DB2 for z/OS Bufferpool Tuning win by Divide and Conquer or Lose by Multiply ...John Campbell
 
DbB 10 Webcast #3 The Secrets Of Scalability
DbB 10 Webcast #3   The Secrets Of ScalabilityDbB 10 Webcast #3   The Secrets Of Scalability
DbB 10 Webcast #3 The Secrets Of ScalabilityLaura Hood
 
Universal Table Spaces for DB2 10 for z/OS - IOD 2010 Seesion 1929 - favero
 Universal Table Spaces for DB2 10 for z/OS - IOD 2010 Seesion 1929 - favero Universal Table Spaces for DB2 10 for z/OS - IOD 2010 Seesion 1929 - favero
Universal Table Spaces for DB2 10 for z/OS - IOD 2010 Seesion 1929 - faveroWillie Favero
 
Efficient Monitoring & Tuning of Dynamic SQL in DB2 for z/OS by Namik Hrle ...
Efficient Monitoring & Tuning of Dynamic SQL in DB2 for z/OS  by  Namik Hrle ...Efficient Monitoring & Tuning of Dynamic SQL in DB2 for z/OS  by  Namik Hrle ...
Efficient Monitoring & Tuning of Dynamic SQL in DB2 for z/OS by Namik Hrle ...Surekha Parekh
 
12 cool features in defrag 12
12 cool features in defrag 1212 cool features in defrag 12
12 cool features in defrag 12aosborne
 
11 cool features in Defrag.nsf+ 11
11 cool features in Defrag.nsf+ 1111 cool features in Defrag.nsf+ 11
11 cool features in Defrag.nsf+ 11aosborne
 
High Availability Options for DB2 Data Centre
High Availability Options for DB2 Data CentreHigh Availability Options for DB2 Data Centre
High Availability Options for DB2 Data Centreterraborealis
 
DB2 LUW - Backup and Recovery
DB2 LUW - Backup and RecoveryDB2 LUW - Backup and Recovery
DB2 LUW - Backup and Recoveryimranasayed
 
Understanding and controlling transaction logs
Understanding and controlling transaction logsUnderstanding and controlling transaction logs
Understanding and controlling transaction logsRed Gate Software
 
Best Practices For Optimizing DB2 Performance Final
Best Practices For Optimizing DB2 Performance FinalBest Practices For Optimizing DB2 Performance Final
Best Practices For Optimizing DB2 Performance FinalDatavail
 

Tendances (16)

DB2 for z/OS Architecture in Nutshell
DB2 for z/OS Architecture in NutshellDB2 for z/OS Architecture in Nutshell
DB2 for z/OS Architecture in Nutshell
 
DB2 Accounting Reporting
DB2  Accounting ReportingDB2  Accounting Reporting
DB2 Accounting Reporting
 
DB2 11 for z/OS Migration Planning and Early Customer Experiences
DB2 11 for z/OS Migration Planning and Early Customer ExperiencesDB2 11 for z/OS Migration Planning and Early Customer Experiences
DB2 11 for z/OS Migration Planning and Early Customer Experiences
 
Db2 recovery IDUG EMEA 2013
Db2 recovery IDUG EMEA 2013Db2 recovery IDUG EMEA 2013
Db2 recovery IDUG EMEA 2013
 
DB2 for z/OS Real Storage Monitoring, Control and Planning
DB2 for z/OS Real Storage Monitoring, Control and PlanningDB2 for z/OS Real Storage Monitoring, Control and Planning
DB2 for z/OS Real Storage Monitoring, Control and Planning
 
DB2 for z/OS Bufferpool Tuning win by Divide and Conquer or Lose by Multiply ...
DB2 for z/OS Bufferpool Tuning win by Divide and Conquer or Lose by Multiply ...DB2 for z/OS Bufferpool Tuning win by Divide and Conquer or Lose by Multiply ...
DB2 for z/OS Bufferpool Tuning win by Divide and Conquer or Lose by Multiply ...
 
DbB 10 Webcast #3 The Secrets Of Scalability
DbB 10 Webcast #3   The Secrets Of ScalabilityDbB 10 Webcast #3   The Secrets Of Scalability
DbB 10 Webcast #3 The Secrets Of Scalability
 
Universal Table Spaces for DB2 10 for z/OS - IOD 2010 Seesion 1929 - favero
 Universal Table Spaces for DB2 10 for z/OS - IOD 2010 Seesion 1929 - favero Universal Table Spaces for DB2 10 for z/OS - IOD 2010 Seesion 1929 - favero
Universal Table Spaces for DB2 10 for z/OS - IOD 2010 Seesion 1929 - favero
 
Efficient Monitoring & Tuning of Dynamic SQL in DB2 for z/OS by Namik Hrle ...
Efficient Monitoring & Tuning of Dynamic SQL in DB2 for z/OS  by  Namik Hrle ...Efficient Monitoring & Tuning of Dynamic SQL in DB2 for z/OS  by  Namik Hrle ...
Efficient Monitoring & Tuning of Dynamic SQL in DB2 for z/OS by Namik Hrle ...
 
12 cool features in defrag 12
12 cool features in defrag 1212 cool features in defrag 12
12 cool features in defrag 12
 
11 cool features in Defrag.nsf+ 11
11 cool features in Defrag.nsf+ 1111 cool features in Defrag.nsf+ 11
11 cool features in Defrag.nsf+ 11
 
High Availability Options for DB2 Data Centre
High Availability Options for DB2 Data CentreHigh Availability Options for DB2 Data Centre
High Availability Options for DB2 Data Centre
 
DB2 LUW - Backup and Recovery
DB2 LUW - Backup and RecoveryDB2 LUW - Backup and Recovery
DB2 LUW - Backup and Recovery
 
Understanding and controlling transaction logs
Understanding and controlling transaction logsUnderstanding and controlling transaction logs
Understanding and controlling transaction logs
 
IMSDB - DBRC
IMSDB - DBRCIMSDB - DBRC
IMSDB - DBRC
 
Best Practices For Optimizing DB2 Performance Final
Best Practices For Optimizing DB2 Performance FinalBest Practices For Optimizing DB2 Performance Final
Best Practices For Optimizing DB2 Performance Final
 

En vedette

Episode 2 Installation Triton Slides
Episode 2 Installation Triton SlidesEpisode 2 Installation Triton Slides
Episode 2 Installation Triton SlidesLaura Hood
 
UKCMG DB2 pureScale
UKCMG DB2 pureScaleUKCMG DB2 pureScale
UKCMG DB2 pureScaleLaura Hood
 
UKGSE DB2 pureScale
UKGSE DB2 pureScaleUKGSE DB2 pureScale
UKGSE DB2 pureScaleLaura Hood
 
DB2 pureScale Technology Preview
DB2 pureScale Technology PreviewDB2 pureScale Technology Preview
DB2 pureScale Technology PreviewCristian Molaro
 
DB2 V 10 HADR Multiple Standby
DB2 V 10 HADR Multiple StandbyDB2 V 10 HADR Multiple Standby
DB2 V 10 HADR Multiple StandbyDale McInnis
 
IBM DB2 LUW UDB DBA Online Training by Etraining.guru
IBM DB2 LUW UDB DBA Online Training by Etraining.guruIBM DB2 LUW UDB DBA Online Training by Etraining.guru
IBM DB2 LUW UDB DBA Online Training by Etraining.guruRavikumar Nandigam
 
Herd your chickens: Ansible for DB2 configuration management
Herd your chickens: Ansible for DB2 configuration managementHerd your chickens: Ansible for DB2 configuration management
Herd your chickens: Ansible for DB2 configuration managementFrederik Engelen
 
Ibm db2 10.5 for linux, unix, and windows upgrading to db2 version 10.5
Ibm db2 10.5 for linux, unix, and windows   upgrading to db2 version 10.5Ibm db2 10.5 for linux, unix, and windows   upgrading to db2 version 10.5
Ibm db2 10.5 for linux, unix, and windows upgrading to db2 version 10.5bupbechanhgmail
 
DBA Basics guide
DBA Basics guideDBA Basics guide
DBA Basics guideazoznasser1
 
Ibm db2 10.5 for linux, unix, and windows what's new for db2 version 10.5
Ibm db2 10.5 for linux, unix, and windows   what's new for db2 version 10.5Ibm db2 10.5 for linux, unix, and windows   what's new for db2 version 10.5
Ibm db2 10.5 for linux, unix, and windows what's new for db2 version 10.5bupbechanhgmail
 
Business Case: IBM DB2 versus Oracle Database - Conor O'Mahony
Business Case: IBM DB2 versus Oracle Database - Conor O'MahonyBusiness Case: IBM DB2 versus Oracle Database - Conor O'Mahony
Business Case: IBM DB2 versus Oracle Database - Conor O'Mahonycomahony
 
Showdown: IBM DB2 versus Oracle Database for OLTP
Showdown: IBM DB2 versus Oracle Database for OLTPShowdown: IBM DB2 versus Oracle Database for OLTP
Showdown: IBM DB2 versus Oracle Database for OLTPcomahony
 
World of Watson - DB2 for Linux, UNIX and Windows Roadmap
World of Watson - DB2 for Linux, UNIX and Windows RoadmapWorld of Watson - DB2 for Linux, UNIX and Windows Roadmap
World of Watson - DB2 for Linux, UNIX and Windows RoadmapIBM_Info_Management
 

En vedette (15)

Episode 2 Installation Triton Slides
Episode 2 Installation Triton SlidesEpisode 2 Installation Triton Slides
Episode 2 Installation Triton Slides
 
UKCMG DB2 pureScale
UKCMG DB2 pureScaleUKCMG DB2 pureScale
UKCMG DB2 pureScale
 
UKGSE DB2 pureScale
UKGSE DB2 pureScaleUKGSE DB2 pureScale
UKGSE DB2 pureScale
 
Db2 v10.5 An Overview
Db2 v10.5 An OverviewDb2 v10.5 An Overview
Db2 v10.5 An Overview
 
DB2 pureScale Technology Preview
DB2 pureScale Technology PreviewDB2 pureScale Technology Preview
DB2 pureScale Technology Preview
 
DB2 V 10 HADR Multiple Standby
DB2 V 10 HADR Multiple StandbyDB2 V 10 HADR Multiple Standby
DB2 V 10 HADR Multiple Standby
 
IBM DB2 LUW UDB DBA Online Training by Etraining.guru
IBM DB2 LUW UDB DBA Online Training by Etraining.guruIBM DB2 LUW UDB DBA Online Training by Etraining.guru
IBM DB2 LUW UDB DBA Online Training by Etraining.guru
 
Herd your chickens: Ansible for DB2 configuration management
Herd your chickens: Ansible for DB2 configuration managementHerd your chickens: Ansible for DB2 configuration management
Herd your chickens: Ansible for DB2 configuration management
 
Ibm db2 10.5 for linux, unix, and windows upgrading to db2 version 10.5
Ibm db2 10.5 for linux, unix, and windows   upgrading to db2 version 10.5Ibm db2 10.5 for linux, unix, and windows   upgrading to db2 version 10.5
Ibm db2 10.5 for linux, unix, and windows upgrading to db2 version 10.5
 
DBA Basics guide
DBA Basics guideDBA Basics guide
DBA Basics guide
 
Ibm db2 10.5 for linux, unix, and windows what's new for db2 version 10.5
Ibm db2 10.5 for linux, unix, and windows   what's new for db2 version 10.5Ibm db2 10.5 for linux, unix, and windows   what's new for db2 version 10.5
Ibm db2 10.5 for linux, unix, and windows what's new for db2 version 10.5
 
Business Case: IBM DB2 versus Oracle Database - Conor O'Mahony
Business Case: IBM DB2 versus Oracle Database - Conor O'MahonyBusiness Case: IBM DB2 versus Oracle Database - Conor O'Mahony
Business Case: IBM DB2 versus Oracle Database - Conor O'Mahony
 
DB2 9.7 Overview
DB2 9.7 OverviewDB2 9.7 Overview
DB2 9.7 Overview
 
Showdown: IBM DB2 versus Oracle Database for OLTP
Showdown: IBM DB2 versus Oracle Database for OLTPShowdown: IBM DB2 versus Oracle Database for OLTP
Showdown: IBM DB2 versus Oracle Database for OLTP
 
World of Watson - DB2 for Linux, UNIX and Windows Roadmap
World of Watson - DB2 for Linux, UNIX and Windows RoadmapWorld of Watson - DB2 for Linux, UNIX and Windows Roadmap
World of Watson - DB2 for Linux, UNIX and Windows Roadmap
 

Similaire à Episode 4 DB2 pureScale Performance Webinar Oct 2010

Getting Under the Hood of Kafka Streams: Optimizing Storage Engines to Tune U...
Getting Under the Hood of Kafka Streams: Optimizing Storage Engines to Tune U...Getting Under the Hood of Kafka Streams: Optimizing Storage Engines to Tune U...
Getting Under the Hood of Kafka Streams: Optimizing Storage Engines to Tune U...HostedbyConfluent
 
Abap interview questions and answers
Abap interview questions and answersAbap interview questions and answers
Abap interview questions and answersKaustav Pyne
 
(ATS6-PLAT06) Maximizing AEP Performance
(ATS6-PLAT06) Maximizing AEP Performance(ATS6-PLAT06) Maximizing AEP Performance
(ATS6-PLAT06) Maximizing AEP PerformanceBIOVIA
 
JourneyToLowCode_2of4.pdf
JourneyToLowCode_2of4.pdfJourneyToLowCode_2of4.pdf
JourneyToLowCode_2of4.pdfVaibhavVaidya30
 
Oracle DB In-Memory technologie v kombinaci s procesorem M7
Oracle DB In-Memory technologie v kombinaci s procesorem M7Oracle DB In-Memory technologie v kombinaci s procesorem M7
Oracle DB In-Memory technologie v kombinaci s procesorem M7MarketingArrowECS_CZ
 
Managing Memory & Locks - Series 1 Memory Management
Managing  Memory & Locks - Series 1 Memory ManagementManaging  Memory & Locks - Series 1 Memory Management
Managing Memory & Locks - Series 1 Memory ManagementDAGEOP LTD
 
SPS Kansas City: What SharePoint Admin need to know about SQL
SPS Kansas City: What SharePoint Admin need to know about SQLSPS Kansas City: What SharePoint Admin need to know about SQL
SPS Kansas City: What SharePoint Admin need to know about SQLJ.D. Wade
 
Troubleshooting SQL Server
Troubleshooting SQL ServerTroubleshooting SQL Server
Troubleshooting SQL ServerStephen Rose
 
Shared Personalization Service - How To Scale to 15K RPS, Patrice Pelland
Shared Personalization Service - How To Scale to 15K RPS, Patrice PellandShared Personalization Service - How To Scale to 15K RPS, Patrice Pelland
Shared Personalization Service - How To Scale to 15K RPS, Patrice PellandFuenteovejuna
 
Ceph Day Berlin: Ceph on All Flash Storage - Breaking Performance Barriers
Ceph Day Berlin: Ceph on All Flash Storage - Breaking Performance BarriersCeph Day Berlin: Ceph on All Flash Storage - Breaking Performance Barriers
Ceph Day Berlin: Ceph on All Flash Storage - Breaking Performance BarriersCeph Community
 
What SQL DBA's need to know about SharePoint-St. Louis 2013
What SQL DBA's need to know about SharePoint-St. Louis 2013What SQL DBA's need to know about SharePoint-St. Louis 2013
What SQL DBA's need to know about SharePoint-St. Louis 2013J.D. Wade
 
071410 sun a_1515_feldman_stephen
071410 sun a_1515_feldman_stephen071410 sun a_1515_feldman_stephen
071410 sun a_1515_feldman_stephenSteve Feldman
 
A Common Database Approach for OLTP and OLAP Using an In-Memory Column Database
A Common Database Approach for OLTP and OLAP Using an In-Memory Column DatabaseA Common Database Approach for OLTP and OLAP Using an In-Memory Column Database
A Common Database Approach for OLTP and OLAP Using an In-Memory Column DatabaseIshara Amarasekera
 
MySQL Cluster Local Checkpoint (LCP) evolution in 7.6 (2019)
MySQL Cluster Local Checkpoint (LCP) evolution in 7.6 (2019)MySQL Cluster Local Checkpoint (LCP) evolution in 7.6 (2019)
MySQL Cluster Local Checkpoint (LCP) evolution in 7.6 (2019)Frazer Clement
 
How to Upgrade Major Version of Your Production PostgreSQL
How to Upgrade Major Version of Your Production PostgreSQLHow to Upgrade Major Version of Your Production PostgreSQL
How to Upgrade Major Version of Your Production PostgreSQLKeisuke Suzuki
 
Finance month closing with HANA
Finance month closing with HANAFinance month closing with HANA
Finance month closing with HANADouglas Bernardini
 
Incremental backups
Incremental backupsIncremental backups
Incremental backupsVlad Lesin
 
SharePoint 2010 database maintenance
SharePoint 2010 database maintenanceSharePoint 2010 database maintenance
SharePoint 2010 database maintenanceMatt Ranlett
 
No C-QL (Or how I learned to stop worrying, and love eventual consistency) (N...
No C-QL (Or how I learned to stop worrying, and love eventual consistency) (N...No C-QL (Or how I learned to stop worrying, and love eventual consistency) (N...
No C-QL (Or how I learned to stop worrying, and love eventual consistency) (N...Brian Brazil
 

Similaire à Episode 4 DB2 pureScale Performance Webinar Oct 2010 (20)

Getting Under the Hood of Kafka Streams: Optimizing Storage Engines to Tune U...
Getting Under the Hood of Kafka Streams: Optimizing Storage Engines to Tune U...Getting Under the Hood of Kafka Streams: Optimizing Storage Engines to Tune U...
Getting Under the Hood of Kafka Streams: Optimizing Storage Engines to Tune U...
 
Abap interview questions and answers
Abap interview questions and answersAbap interview questions and answers
Abap interview questions and answers
 
(ATS6-PLAT06) Maximizing AEP Performance
(ATS6-PLAT06) Maximizing AEP Performance(ATS6-PLAT06) Maximizing AEP Performance
(ATS6-PLAT06) Maximizing AEP Performance
 
JourneyToLowCode_2of4.pdf
JourneyToLowCode_2of4.pdfJourneyToLowCode_2of4.pdf
JourneyToLowCode_2of4.pdf
 
Oracle DB In-Memory technologie v kombinaci s procesorem M7
Oracle DB In-Memory technologie v kombinaci s procesorem M7Oracle DB In-Memory technologie v kombinaci s procesorem M7
Oracle DB In-Memory technologie v kombinaci s procesorem M7
 
Managing Memory & Locks - Series 1 Memory Management
Managing  Memory & Locks - Series 1 Memory ManagementManaging  Memory & Locks - Series 1 Memory Management
Managing Memory & Locks - Series 1 Memory Management
 
SPS Kansas City: What SharePoint Admin need to know about SQL
SPS Kansas City: What SharePoint Admin need to know about SQLSPS Kansas City: What SharePoint Admin need to know about SQL
SPS Kansas City: What SharePoint Admin need to know about SQL
 
Troubleshooting SQL Server
Troubleshooting SQL ServerTroubleshooting SQL Server
Troubleshooting SQL Server
 
Shared Personalization Service - How To Scale to 15K RPS, Patrice Pelland
Shared Personalization Service - How To Scale to 15K RPS, Patrice PellandShared Personalization Service - How To Scale to 15K RPS, Patrice Pelland
Shared Personalization Service - How To Scale to 15K RPS, Patrice Pelland
 
Ceph Day Berlin: Ceph on All Flash Storage - Breaking Performance Barriers
Ceph Day Berlin: Ceph on All Flash Storage - Breaking Performance BarriersCeph Day Berlin: Ceph on All Flash Storage - Breaking Performance Barriers
Ceph Day Berlin: Ceph on All Flash Storage - Breaking Performance Barriers
 
What SQL DBA's need to know about SharePoint-St. Louis 2013
What SQL DBA's need to know about SharePoint-St. Louis 2013What SQL DBA's need to know about SharePoint-St. Louis 2013
What SQL DBA's need to know about SharePoint-St. Louis 2013
 
071410 sun a_1515_feldman_stephen
071410 sun a_1515_feldman_stephen071410 sun a_1515_feldman_stephen
071410 sun a_1515_feldman_stephen
 
A Common Database Approach for OLTP and OLAP Using an In-Memory Column Database
A Common Database Approach for OLTP and OLAP Using an In-Memory Column DatabaseA Common Database Approach for OLTP and OLAP Using an In-Memory Column Database
A Common Database Approach for OLTP and OLAP Using an In-Memory Column Database
 
MySQL Cluster Local Checkpoint (LCP) evolution in 7.6 (2019)
MySQL Cluster Local Checkpoint (LCP) evolution in 7.6 (2019)MySQL Cluster Local Checkpoint (LCP) evolution in 7.6 (2019)
MySQL Cluster Local Checkpoint (LCP) evolution in 7.6 (2019)
 
How to Upgrade Major Version of Your Production PostgreSQL
How to Upgrade Major Version of Your Production PostgreSQLHow to Upgrade Major Version of Your Production PostgreSQL
How to Upgrade Major Version of Your Production PostgreSQL
 
Finance month closing with HANA
Finance month closing with HANAFinance month closing with HANA
Finance month closing with HANA
 
Kafka internals
Kafka internalsKafka internals
Kafka internals
 
Incremental backups
Incremental backupsIncremental backups
Incremental backups
 
SharePoint 2010 database maintenance
SharePoint 2010 database maintenanceSharePoint 2010 database maintenance
SharePoint 2010 database maintenance
 
No C-QL (Or how I learned to stop worrying, and love eventual consistency) (N...
No C-QL (Or how I learned to stop worrying, and love eventual consistency) (N...No C-QL (Or how I learned to stop worrying, and love eventual consistency) (N...
No C-QL (Or how I learned to stop worrying, and love eventual consistency) (N...
 

Plus de Laura Hood

Top 10 DB2 Support Nightmares #10
Top 10 DB2 Support Nightmares  #10Top 10 DB2 Support Nightmares  #10
Top 10 DB2 Support Nightmares #10Laura Hood
 
Top 10 DB2 Support Nightmares #9
Top 10 DB2 Support Nightmares  #9Top 10 DB2 Support Nightmares  #9
Top 10 DB2 Support Nightmares #9Laura Hood
 
Top 10 DB2 Support Nightmares #8
Top 10 DB2 Support Nightmares  #8Top 10 DB2 Support Nightmares  #8
Top 10 DB2 Support Nightmares #8Laura Hood
 
Top 10 DB2 Support Nightmares #7
Top 10 DB2 Support Nightmares  #7 Top 10 DB2 Support Nightmares  #7
Top 10 DB2 Support Nightmares #7 Laura Hood
 
Top 10 db2 support nightmares #6
Top 10 db2 support nightmares #6Top 10 db2 support nightmares #6
Top 10 db2 support nightmares #6Laura Hood
 
Consultancy on Demand - Infographic
Consultancy on Demand - InfographicConsultancy on Demand - Infographic
Consultancy on Demand - InfographicLaura Hood
 
A Time Traveller's Guide to DB2: Technology Themes for 2014 and Beyond
A Time Traveller's Guide to DB2: Technology Themes for 2014 and BeyondA Time Traveller's Guide to DB2: Technology Themes for 2014 and Beyond
A Time Traveller's Guide to DB2: Technology Themes for 2014 and BeyondLaura Hood
 
Top 10 DB2 Support Nightmares #1
Top 10 DB2 Support Nightmares  #1Top 10 DB2 Support Nightmares  #1
Top 10 DB2 Support Nightmares #1Laura Hood
 
Db2 10 memory management uk db2 user group june 2013 [read-only]
Db2 10 memory management   uk db2 user group june 2013 [read-only]Db2 10 memory management   uk db2 user group june 2013 [read-only]
Db2 10 memory management uk db2 user group june 2013 [read-only]Laura Hood
 
DB2 10 Security Enhancements
DB2 10 Security EnhancementsDB2 10 Security Enhancements
DB2 10 Security EnhancementsLaura Hood
 
DB2 10 Webcast #2 - Justifying The Upgrade
DB2 10 Webcast #2  - Justifying The UpgradeDB2 10 Webcast #2  - Justifying The Upgrade
DB2 10 Webcast #2 - Justifying The UpgradeLaura Hood
 
DB2 10 Webcast #1 - Overview And Migration Planning
DB2 10 Webcast #1 - Overview And Migration PlanningDB2 10 Webcast #1 - Overview And Migration Planning
DB2 10 Webcast #1 - Overview And Migration PlanningLaura Hood
 
Time Travelling With DB2 10 For zOS
Time Travelling With DB2 10 For zOSTime Travelling With DB2 10 For zOS
Time Travelling With DB2 10 For zOSLaura Hood
 
DB2DART - DB2Night Show October 2011
DB2DART - DB2Night Show October 2011DB2DART - DB2Night Show October 2011
DB2DART - DB2Night Show October 2011Laura Hood
 
DB2 z/OS & Java - What\'s New?
DB2 z/OS & Java - What\'s New?DB2 z/OS & Java - What\'s New?
DB2 z/OS & Java - What\'s New?Laura Hood
 
Temporal And Other DB2 10 For Z Os Highlights
Temporal And Other DB2 10 For Z Os HighlightsTemporal And Other DB2 10 For Z Os Highlights
Temporal And Other DB2 10 For Z Os HighlightsLaura Hood
 
DB210 Smarter Database IBM Tech Forum 2011
DB210 Smarter Database   IBM Tech Forum 2011DB210 Smarter Database   IBM Tech Forum 2011
DB210 Smarter Database IBM Tech Forum 2011Laura Hood
 
Managing The Data Explosion
Managing The Data ExplosionManaging The Data Explosion
Managing The Data ExplosionLaura Hood
 
Managing the financial services data explosion
Managing the financial services data explosionManaging the financial services data explosion
Managing the financial services data explosionLaura Hood
 

Plus de Laura Hood (19)

Top 10 DB2 Support Nightmares #10
Top 10 DB2 Support Nightmares  #10Top 10 DB2 Support Nightmares  #10
Top 10 DB2 Support Nightmares #10
 
Top 10 DB2 Support Nightmares #9
Top 10 DB2 Support Nightmares  #9Top 10 DB2 Support Nightmares  #9
Top 10 DB2 Support Nightmares #9
 
Top 10 DB2 Support Nightmares #8
Top 10 DB2 Support Nightmares  #8Top 10 DB2 Support Nightmares  #8
Top 10 DB2 Support Nightmares #8
 
Top 10 DB2 Support Nightmares #7
Top 10 DB2 Support Nightmares  #7 Top 10 DB2 Support Nightmares  #7
Top 10 DB2 Support Nightmares #7
 
Top 10 db2 support nightmares #6
Top 10 db2 support nightmares #6Top 10 db2 support nightmares #6
Top 10 db2 support nightmares #6
 
Consultancy on Demand - Infographic
Consultancy on Demand - InfographicConsultancy on Demand - Infographic
Consultancy on Demand - Infographic
 
A Time Traveller's Guide to DB2: Technology Themes for 2014 and Beyond
A Time Traveller's Guide to DB2: Technology Themes for 2014 and BeyondA Time Traveller's Guide to DB2: Technology Themes for 2014 and Beyond
A Time Traveller's Guide to DB2: Technology Themes for 2014 and Beyond
 
Top 10 DB2 Support Nightmares #1
Top 10 DB2 Support Nightmares  #1Top 10 DB2 Support Nightmares  #1
Top 10 DB2 Support Nightmares #1
 
Db2 10 memory management uk db2 user group june 2013 [read-only]
Db2 10 memory management   uk db2 user group june 2013 [read-only]Db2 10 memory management   uk db2 user group june 2013 [read-only]
Db2 10 memory management uk db2 user group june 2013 [read-only]
 
DB2 10 Security Enhancements
DB2 10 Security EnhancementsDB2 10 Security Enhancements
DB2 10 Security Enhancements
 
DB2 10 Webcast #2 - Justifying The Upgrade
DB2 10 Webcast #2  - Justifying The UpgradeDB2 10 Webcast #2  - Justifying The Upgrade
DB2 10 Webcast #2 - Justifying The Upgrade
 
DB2 10 Webcast #1 - Overview And Migration Planning
DB2 10 Webcast #1 - Overview And Migration PlanningDB2 10 Webcast #1 - Overview And Migration Planning
DB2 10 Webcast #1 - Overview And Migration Planning
 
Time Travelling With DB2 10 For zOS
Time Travelling With DB2 10 For zOSTime Travelling With DB2 10 For zOS
Time Travelling With DB2 10 For zOS
 
DB2DART - DB2Night Show October 2011
DB2DART - DB2Night Show October 2011DB2DART - DB2Night Show October 2011
DB2DART - DB2Night Show October 2011
 
DB2 z/OS & Java - What\'s New?
DB2 z/OS & Java - What\'s New?DB2 z/OS & Java - What\'s New?
DB2 z/OS & Java - What\'s New?
 
Temporal And Other DB2 10 For Z Os Highlights
Temporal And Other DB2 10 For Z Os HighlightsTemporal And Other DB2 10 For Z Os Highlights
Temporal And Other DB2 10 For Z Os Highlights
 
DB210 Smarter Database IBM Tech Forum 2011
DB210 Smarter Database   IBM Tech Forum 2011DB210 Smarter Database   IBM Tech Forum 2011
DB210 Smarter Database IBM Tech Forum 2011
 
Managing The Data Explosion
Managing The Data ExplosionManaging The Data Explosion
Managing The Data Explosion
 
Managing the financial services data explosion
Managing the financial services data explosionManaging the financial services data explosion
Managing the financial services data explosion
 

Episode 4 DB2 pureScale Performance Webinar Oct 2010

  • 1. DB2 pureScale Performance Steve Rees srees@ca.ibm.com Oct 19, 2010
  • 2. 2 Copyright IBM 2010 Agenda DB2 pureScale technology review RDMA and low-latency interconnect Monitoring and tuning bufferpools in pureScale Architectural features for top performance Performance metrics
  • 3. 3 Copyright IBM 2010 Cluster Interconnect DB2 pureScale : Technology Review Single Database View Clients Database Log Log Log Log Shared Storage Access CS CS CSCS CS CS CS Member Member Member Member Primary2nd-ary DB2 engine runs on several host computers – Co-operate with each other to provide coherent access to the database from any member Data sharing architecture – Shared access to database – Members write to their own logs – Logs accessible from another host (used during recovery) Cluster Caching Facility (CF) technology from STG – Efficient global locking and buffer management – Synchronous duplexing to secondary ensures availability Low latency, high speed interconnect – Special optimizations provide significant advantages on RDMA- capable interconnects like Infiniband Clients connect anywhere,… … see single database – Clients connect into any member – Automatic load balancing and client reroute may change underlying physical member to which client is connected Integrated cluster services – Failure detection, recovery automation, cluster file system – In partnership with STG (GPFS,RSCT) and Tivoli (SA MP) Leverage IBM’s System z Sysplex Experience and Know-How
  • 4. 4 Copyright IBM 2010 DB2 pureScale and low-latency interconnect Infiniband & uDAPL provide the low-latency RDMA infrastructure exploited by pureScale pureScale currently uses DDR and QDR IB adapters according to platform – Peak throughput of about 2-4 M messages per second – Provide message latencies in the 10s of microseconds or even lower The Infiniband development roadmap indicates continued increases in bit rates Infiniband Roadmap from www.infinibandta.org
  • 5. 5 Copyright IBM 2010 Two-level page buffering – data consistency & improved performance The local bufferpool (LBP) caches both read-only and updated pages for that member The shared GBP contains references to every page in all LBPs across the cluster – References ensure consistency across members – who’s interested in which pages, in case the pages are updated The GBP also contains copies of all updated pages from the LBPs – Sent from the LBP at transaction commit time – Stored in the GBP & available to members on demand – 30 µs page read request over Infiniband from the GBP can be more than 100x faster than reading from disk Statistics are kept for tuning – Found in LBP vs. found in GBP vs. read from disk – Useful in tuning GBP / LBP sizes CF M1 M2 M3 10µs 3 5000µs 60µs 30 µs 2 30µs 1 Expensive disk reads from M1, M2 not required – get the modified page from the CF #1 #2 #3 #4#5
  • 6. 6 Copyright IBM 2010 pureScale bufferpool monitoring and tuning Familiar DB2 hit ratio calculations are useful with pureScale – HR = (logical reads – physical reads) / logical reads e.g. (pool_data_l_reads – pool_data_p_reads)/pool_data_l_reads – As usual, physical reads come from disk, logical reads from the bufferpool (in pureScale, either this means either the LBP or GBP) e.g., pool_data_l_reads = pool_data_lbp_pages_found + pool_data_gbp_l_reads New metrics in pureScale support breaking this down by LBP & GBP amounts – pool_data_lbp_pages_found = logical data reads satisfied by the LBP • i.e., we needed a page, and it was present & valid in the LBP – pool_data_gbp_l_reads = logical data reads attempted at the GBP • i.e., either not present or not valid in the LBP, so we needed to go to the GBP – pool_data_gbp_p_reads = physical data reads due to page not present in either the LBP or GBP • Essentially the same as non-pureScale pool_data_p_reads – pool_data_gbp_invalid_pages = number of GBP data read attempts due to an LBP page being present but marked invalid • An indicator of the rate of GBP updates & their impact on the LBP Of course, there are index ones too
  • 7. 7 Copyright IBM 2010 pureScale bufferpool monitoring Overall (and non-pureScale) hit ratio – (pool_data_l_reads – pool_data_p_reads)/pool_data_l_reads – Great values: 95% for index, 90% for data – Good values: 80-90% for index, 75-85% for data LBP hit ratio – (pool_data_lbp_pages_found / pool_data_l_reads) * 100% – Generally lower than the overall hit ratio, since it excludes GBP hits – Factors which may affect it, other than LBP size • Increases with greater portion of read activity in the system – Decreasing probability that LBP copies of the page have been invalidated • May decrease with cluster size – Increasing probability that another member has invalidated the LBP page GBP hit ratio – (pool_data_gbp_l_reads – pool_data_gbp_p_reads) / pool_data_gbp_l_reads – A hit here is a read of a previously modified page, so hit ratios are typically quite low • An overall (LBP+GBP) H/R in the high 90's can correspond to a GBP H/R in the low 80's – Factors which may affect it, other than GBP size • Decreases with greater portion of read activity
  • 8. 8 Copyright IBM 2010 pureScale bufferpool tuning Step 1: typical rule-of-thumb for GBP size = 35-40% of Σ( all members’ LBP sizes ) e.g. 4 members, LBP size of 1M pages each -> GBP size of 1.4 to 1.6M pages NB - don't forget, GBP page size is always 4kB, no matter what the LBP page size is. – If your workload very read-heavy (e.g. 90% read), initial GBP allocation could be in the 20-30% range – For 2-member clusters, you may want to start with 40-50% of total LBP, vs. 35-40% Step 2: monitor the overall BP hit ratio as usual, with pool_data_l_reads, pool_data_p_reads, etc. – Meets your goals? If yes, then done! Step 3: check LBP H/R with pool_data_lbp_pages_found/pool_data_l_reads – Great values: 90% for index, 85% for data – Good values: 70-80% for index, 65-80% for data – Increasing LBP size can help increase LBP H/R – NB – for each 16 extra LBP pages, the GBP needs 1 extra page for registrations Step 4: check GBP H/R with pool_data_gbp_l_reads, pool_data_gbp_p_reads, etc. – Great values: 90% for index, 80% for data – Good values: 65-80% for index, 60-75% for data – pool_data_l_reads > 10 x pool_data_gbp_l_reads means low GBP dependence – may mean tuning GBP size in this case is less valuable – pool_data_gbp_invalid_pages > 25% of pool_data_gbp_l_reads means GBP is really helping out, and could benefit from extra pages
  • 9. 9 Copyright IBM 2010 Page lock negotiation – or Psst! Hey buddy, can you pass me that page? – pureScale page locks are physical locks, indicating which member currently ‘owns’ the page. Picture the following: • Member A : acquires a page P and modifies a row on it, and continues with its transaction. ‘A’ holds an exclusive page lock on page P until ‘A’ commits • Member B : wants to modify a different row on the same page P. What now? – ‘B’ doesn’t have to wait until ‘A’ commits & releases the page lock • The CF will negotiate the page back from ‘A’ in the middle of ‘A’s transaction, on ‘B’s behalf • Provides far better concurrency & performance than needing to wait for a page lock until the holder commits. Log P P pureScale architectural features for optimum performance P P Member A Member B Log P ?P ! CF GLM Px : A: B
  • 10. 10 Copyright IBM 2010 pureScale architectural features for optimum performance Table append cache and index page cache – What happens in the case of rapid inserts into a single table by multiple members? Or rapid index updates? Will it cause the insert page to ‘thrash’ back & forth between the members, each time one has a new row? – No - each member sets aside an extent for insertion into the table to eliminate contention & page thrashing. Similarly for indexes with the page cache Lock avoidance – pureScale exploits cursor stability (CS) locking semantics to avoid taking locks in many common cases – Reduces pathlength and saves trips to the CF – Transparent & always on
  • 11. 11 Copyright IBM 2010 Notes on storage configuration for performance GPFS best practices – Automatically configured by db2cluster command • Blocksize >= 1 MB (vs. default 64k) provides noticeably improved performance • Direct (unbuffered) IO for both logs & tablespace containers • SCSI-3 P/R on AIX enables faster disk takeover on member failure – Separate paths for logs & tablespaces are recommended Dominant storage performance factor for pureScale: fast log writes – Always important in OLTP – Extra important in pureScale due to log flushes driven by page reclaims – Separate filesystems, separate devices from each other & from tablespaces – Ideally – comfortably under 1ms – Possibly even SSDs to keep write latencies as low as possible
  • 12. 12 Copyright IBM 2010 12 Member Scalability Example Moderately heavy transaction processing workload modeling warehouse & ordering process – Write transactions rate 20% – Typical read/write ratio of many OLTP workloads No cluster awareness in the app – No affinity – No partitioning – No routing of transactions to members Configuration – Twelve 8-core p550 members, 64 GB, 5 GHz – IBM 20Gb/s IB HCAs + 7874-024 IB Switch – Duplexed PowerHA pureScale across 2 additional 8-core p550s, 64 GB, 5 GHz – DS8300 storage 576 15K disks, Two 4Gb FC Switches 1Gb Ethernet Client Connectivity 20Gb IB pureScale Interconnect 7874-024 Switch Two 4Gb FC Switches DS8300 Storage p550 members p550 Cluster Caching Facility Clients (2-way x345)
  • 13. 13 Copyright IBM 2010 12 Member Scalability Example - Results 0 1 2 3 4 5 6 7 8 9 10 11 12 0 5 10 15 1.98x @ 2 members 3.9x @ 4 members # Members Throughputvs.1member 7.6x @ 8 members 10.4x @ 12 members
  • 14. 14 Copyright IBM 2010 DB2 pureScale Architecture Scalability How far will it scale? Take a web commerce type workload – Read mostly but not read only – about 90/10 Don’t make the application cluster aware – No routing of transactions to members – Demonstrate transparent application scaling Scale out to the 128 member limit and measure scalability
  • 15. 15 Copyright IBM 2010 The 128-member result 64 Members 95% Scalability 16 Members Over 95% Scalability 2, 4 and 8 Members Over 95% Scalability 32 Members Over 95% Scalability 88 Members 90% Scalability 112 Members 89% Scalability 128 Members 84% Scalability
  • 16. 16 Copyright IBM 2010 Summary Performance & scalability are two top goals of pureScale – many architectural features were designed solely to drive the best possible performance Monitoring and tuning for pureScale extends existing DB2 interfaces and practices – e.g., techniques for optimizing GBP/LBP configuration builds on steps already familiar to DB2 DBAs. The pureScale architecture exploits leading-edge low latency interconnects and RDMA, to achieve excellent performance & scalability – Initial 12- & 128-member proofpoints are strong evidence of a successful first release, with even better things to come!